It always reminds of the story of the creator of counter strike. Every new release he would get a ton of complaints from players about things they didn't even change. Notably that each version had more lag. And he got so fed that he start to negatively subtract peoples pings. And suddenly a ton of players reported back that the change was incredibly good.

Point is, I really don't buy all the stories about a model suddenly being downgraded without at least a modicum of substance. People are grasping at straws in the noise.

$40/hour is $83,000 per year, that is more then livable in the US, and by definition of the median half make more. So I can't agree that it's not a living wage for most.

Assuming they work 40 hours a week. Do they?

Maybe I’m just off the mark but this doesn’t seem like the kind of job you clock in at 9 and stay till 5.

I would expect much more hours when performing and less when not. I would expect it indeed would be approximately a fulltime job.

it's nowhere close. Most orchestras are playing 1-2 nights a week (max) and each performance has 0 or 1 rehearsal (1 if it's a new program, 0 otherwise). Average when performing is maybe 20hr a week

I don't think "destroy humanity" is in the eye of the beholder. The most clear interpretation is the extinction of the human race.

Well if you go by total extinction, then even Skynet in Terminator doesn't count given that there were people left over to resist.

I don't think it's impossible to upset the balance of value in a Mansa Musa kind of way that can lead to black death levels of destruction though resource misallocation. Unlikely, sure. But with the wrong kind of people in the wrong place? Could end up pretty bad. We've built our society as a great filter that funnels sociopaths and psychopaths to the very top by selecting for lack of empathy, and now it's primed and ready to bite us in the ass.

Indeed at the time of skynet in terminator humanity was not destroyed yet. But it was on a clear path to destruction.

mindless idiots feel no need to procreate eventually due to brain rewards being fully hijcked. humanity destroyed in a few generations

As far as we can see in various societies it's actually the highly educated and prosperous people who feel no need to procreate.

I dont think it will happen. AI models are kneecapped. Only a tiny tiny tiny fraction of people are on the list of even being able to use these tools for such things.

Even in a world where these models are heavily restricted, surely the likes of cancer researchers will be among those who have access

Sub-agents. I have 7 20x accounts and I burn them within 1-2 days if I go fully parallel. In some scenarios I use 50 sub-agents for a session which is literally hours of usage for a single 20x account. I'm at the point where I need to parallelize over multiple machines because I just don't have enough CPU and RAM.

What are you doing with them you need so many? It sounds like a Gas Town situation, that you invented an exponential token burning machine.

Decompilation of a game and another larger decompile project. I'm working on it solo. I use 50 sub-agent, one per target function or translation unit. Often there is some progress in a unit but it's not done. So it requires a lot of cycles per function. Notably a single ~80kb function took about a week of constant sol-ultra attention before reaching exactness. The game I'm targeting has ~5000 total functions. The other decompile project has ~10k+ functions.

I'm sure I could be more token efficient, but this was/is also a learning process for me since I never did such an extremely large project before that would take multiple man years before AI.

Fascinating! I think that’s the main difference is my usage is probably tool-bound, meaning it writes some code but then there’s a long period of verification where it compiles things and then waits for the compilation and CI to complete before it can continue. That probably doesn’t consume as many tokens as constantly churning on a problem despite the same wall time.

Yes, this is why I mentioned having so many parallel agents and being compute bound. I run on my own laptop and 2 high-end desktop machines all with 64gb RAM. And it still occasionally happens that one OOM kills codex. They also mostly run unattended until I need to switch their accounts because a usage limit has been hit. Each instance usually can keep going when I sleep or do other things.

I only save the last 30% of usage on a single account for most of my other work, and that is almost always enough.

Sounds like you might benefit from running a custom harness then, no? I can't imagine for a task such as that- that codex is the best option.

Are you not better off with cheaper models like GLM 5.3 or GLM 5.3 Flash? As a lot of that work is repetitive and only needs a stronger model at later stages of cleanup, no?

Of course I don't run everything on Sol. But weaker models can only do very basic things. I have a round robin style workflow where each unit first gets 10 cycles of each luna medium -> high -> terra medium -> terra high -> Sol Medium -> Sol high -> Sol Ultra. And you can clearly see that luna and terra only get a small amount of real progress.

The only subagents I use are Luna.

what are you doing with that many agents/tokens? very curious

See my other comment on your sibling that asked the same.

Are we living in the same universe? Musk and Altman are glossed over and not criticized?

Criticized as being poor human beings, or as being poor executives?

Because the former, as P.T. Barnum pointed out, is just free advertising. And I see far more of that than the latter. Now, if and when their various empires collapse there will be no shortage of people pointing back and saying "all the signs were there", but in my everyday life (obviously just a single point of anecdata), it feels like I see a lot more coverage of them as "bad people" than as "bad at what they claim to be experts on".

I don't see any sentiment in that data.

Would you like a list of pro Musk and Altman users based on their comment and submission history? I'm unsure if that violates the site guidelines, and I prefer not to have my subthreads detached or stern reprimands by mods (it makes me feel bothersome in a host's home).

Profiling Hacker News users based on their comments - https://news.ycombinator.com/item?id=47473086 - March 2026 (86 comments)

(simonw's work I cribbed off of)

I certainly believe there are some Musk and Altman glazers on this site. There are also certainly Ed Zitron glazers.

The point is not a few users you can point out, it wouldn't prove anything. The point is that you claimed that Musk and Altman are "Pathological liars like Musk and Altman are glossed over". I would say that requires at least sentiment analysis that an extremely high percentage(90+%) of hackernews users match that description. And that is simply not what I observe when I use this site. My feeling is that sentiment for Musk and Altman are neutral at best.

I will return with data.

This is really it imo. Fable 5 is better then Sol. But Fable is just of the table for anything even remotely long running. Unless you have very deep pockets. And the difference between Fable and Sol is not world shattering if you ask me. I also find codex a ton better than claude.

So when do you use Fable? For difficult singular tasks?

Yes, I use it when Sol Ultra fails to find a solution.

Its really great how we can breathe new life into older games like this. I'm currently working solo on a Tales of Symphonia gamecube decomp. My aim is to let AI do everything of substance. currently about 3400 functions are exact out of about 4800. So still a way to go. But I haven't written a single line yet.

How did you get started? I have a project in mind and would love to try, but getting a foothold on it feels really hard when you are starting with an ISO image and an empty folder.

For gamecube games you are in luck because there is a template: https://github.com/encounter/dtk-template

But in essence it's not so complex. You only need the original compiler that was used to compile the game and tools like ghidra. Binutils for the architecture can help. From there you ask the agent to start building the project. Prepare to spend a ton of tokens though if you go the full AI route.

Thank you for the template! This worked! It turns out Claude seems to already understand quite a bit about decomps, and if you can find a game or two that already uses the toolchain that your target game uses, it can crib a lot of the basics like MSL and Dolphin/OS pretty quickly. I have barely scratched the surface of the game code, though. I'm trying to actually figure out what things are as we go, and you weren't kidding about the rate of token spend.

ya with Tales of Symphonia I'm currently at about 250 Billion tokens using codex. The multiple resets you get on codex are really helping. It wouldn't surprise me if I crack the trillion tokens for it. I did waste a lot tokens on tooling cycles that went nowhere.

The agents seem to work a lot better if you keep tooling minimal and let it decide ad-hoc.

I have no idea what my token counts are, but I ended up upgrading Claude to the $100 plan, and I've still hit the 5-hour token exhaustion limit twice today. I consider this a good thing--if I'm paying for the tokens, I want to use them! It needs a ton of steering or it will just spin in circles and get nothing done, but if you can lead it towards interesting problems, it just chews through the work. It sure makes a mess though: lots of .c files with one tiny function and PAGES of comments.

TBH I'm not even that interested in getting byte-for-byte reconstruction. I just want to bring an old game back from the dead.

Yeah I get that. The problem is that with AI decomps you really need a clear and verifiable target and exactness solves that problem. If you step away from that you basically leave any agent free space to cheat.

Yeah, that makes sense. Even taking AI out of the equation, that's how you know a decomp is actually right--it matches.

Sadly, even that isn't really a guarantee. I've seen some of the crazy nonsense that the permuter does to get a 100% match. `if ((!x) && (!x) && (!x)) {}` apparently gets different register allocation than `if (!x) {}`, for example. And sometimes important code takes zero instructions (like forwarding an argument from function X to function Y).

This is exactly why I run asahi. There really is not a laptop that comes close to the hardware quality of an apple laptop. I have even sacrificed x86 at the altar which is a much bigger step than battery life imo.

I could not disagree more. I have been using linux for about 15 years. Linux is better now than it ever was.

I've been using Linux as myninly desktop OS for ~25 years now and I agree. I can't really understand the irrational hate of a vocal minority towards systemd, pulseaudio etc

I think there has always been a gate keeping minority that liked the dysfunction. Or more specifically, they liked that they could figure out how to correct the dysfunction while others struggled. That really obnoxious, insecure, computer nerd personality type. So glad I can just block those motherfuckers now instead of begging for their assistance.

Exactly. I have forgot at this point how things use to be with how well everything just works.

I am typing this on a gaming laptop with a Nvidia card that probably would have been a nightmare trying to use linux not too long ago but instead it is an absolute dream machine.

To me, Linux was at its peak around 2005-2010. From then on it went downhill, unfortunately.

That's a baffling take. The era when hardware compatibility was a coin flip (I vividly remember having to download WiFi adapter drivers on another computer to get a fresh Linux install connected), when most desktop environments were poorly developed and very prone to crashing, was your favourite? I suspect this might be a big case of rose-tinted glasses. Linux has never been more "it just works" and accessible than it is now.